R/GPCA.R

In RMAWGEN: Multi-Site Auto-Regressive Weather GENerator

NULL
#'
#' This function makes a Gaussianization procedure based on PCA iteration ( see \code{\link{GPCA_iteration}})
#'
#' @param x_prev previous set of the random variable \code{x}. If it is a \code{varest} object, the residuals are taken into account. 
#' @param n number of reiterations
#' @param extremes  see \code{\link{normalizeGaussian_severalstations}}
#' @author Emanuele Cordano 
#' 
#' @export
#' 
#' @return A  \code{\link{GPCA-class}} S3 object returned by \code{\link{GPCA_iteration}} at each iteration 
#' 			and the final results of the G-PCA procedure (matrix \code{final_results})
#' 
#' @seealso \code{\link{GPCA}},\code{\link{GPCA_iteration}},\code{\link{inv_GPCA_iteration}},\code{\link{inv_GPCA}},\code{\link{GPCA-class}} for 'GPCA' S3 class
#' 
#' 
#' @note This function re-iterates the equation (1) of "PCA Gaussianization for One-Class Remote Sensing Image" by V. Laparra et al., \url{http://www.uv.es/lapeva/papers/SPIE09_one_class.pdf},\url{http://www.uv.es/vista/vistavalencia/papers/SPIE_09_Gaussianization_presentation.pdf} 
##### \url{http://dx.doi.org/doi/10.1117/12.834011} 
#' @examples 
#' library(RMAWGEN)
#' set.seed(1222)
#' nIterations <- 30
#' N <- 20
#' x <- rexp(N)
#' y <- x+rnorm(N)
#' df <- data.frame(x=x,y=y)
#' 
#' GPCA <- GPCA(df,n=nIterations,extremes=TRUE)
#' 
#' x <- rnorm(N)
#' y <- x+rnorm(N)
#' dfn <- data.frame(x=x,y=y)
#' 
#' GPCAn <- GPCA(dfn,n=nIterations,extremes=TRUE)
#' 
#' 




GPCA <- function (x_prev,n=30,extremes=TRUE) {
	
	if ((class(x_prev)=='varest2')  | (class(x_prev)=='GPCAvarest2') | (class(x_prev)=='varest') ) {
		
		x_prev <- residuals(x_prev@VAR)
		
	}
	
	out <- list()
	
	if (n<1) {
		class(out) <- "GPCA"	
		return(out)
	}
	for (i in 1:n) {
		DT <- GPCA_iteration(x_prev=x_prev,extremes=extremes)
		x_prev <- DT$x_next
		out[[i]] <- DT 
	}
	
	iter_names <- sprintf("iteraton%03d",1:length(out))
	
	out[[length(out)+1]] <- normalizeGaussian_severalstations(x_prev,data=x_prev,extremes=extremes,inverse=FALSE) 
	
		
	names(out) <- c(iter_names,"final_results")

	class(out) <- "GPCA"
	return(out)

}